Polyethylene glycol/polylactic acid block co-polymers as solid–solid phase change materials

Abstract

Phase change materials (PCMs) are promising thermal energy storage materials due to their high specific latent heat. Conventional PCMs typically exploit the solid–liquid (s–l) transition. However, leakage and leaching are common issues for solid–liquid PCMs, which have to be addressed before usage in practical applications. In contrast, solid–solid (s–s) PCMs would naturally overcome these issues due to their inherent form stability and homogeneity. In this study, we report a new type of s–s PCM based on chemically linked polyethylene glycol (PEG, the PCM portion) with polylactic acid (PLA, the support portion) in the form of a block co-polymer. Solid-solid latent heat of up to 56 J/g could be achieved, with melting points of between 44 °C and 55 °C. For comparison, PEG was physically mixed into a PLA matrix to form a PEG:PLA composite. However, the composite material saw leakage of up to 9% upon heating, with a corresponding loss in thermal storage capacity. In contrast, the mPEG/PLA block co-polymers were found to be completely homogeneous and thermally stable even when heated above its phase transition temperature, with no observable leakage, demonstrating the superiority of chemical linking strategies in ensuring form stability.

INTRODUCTION

Phase change materials (PCMs) are important components of thermal management systems in various different industries such as logistics,¹ construction,^2–4 and consumer electronics.^5,6 The most commonly employed PCMs are by far are solid–liquid (s–l) PCMs such as paraffin and polyethylene glycol (PEG),^7–9 due to their high latent heat, and tunable melting point ranges.^10–12 However, by virtue of its phase transition, s–l PCMs tend to face problems associated with PCM leakages, especially over the long term. Much research has gone into overcoming this challenge. Broadly speaking, the strategies employed can be classified as encapsulation-based strategies,^13–17 and matrix containment strategies.^18–22 These strategies have met with significant success, with many matrix achieving 60%–80% PCM loading, with a corresponding thermal energy storage capacity of more than 120 J/g being fairly commonplace.²³ However, some fundamental challenges remain for these strategies. In particular, encapsulation-based strategies face the issue of shell cracking, resulting in PCM leakages, while matrix containment strategies tend to be based on weak interactions such as physical adsorption, which would also result in leaching over the long term. This can be highly undesirable in applications where PCM contamination is a significant issue, such as in circuit boards, electronic components, logistic transportation, or even concrete structural strength. Furthermore, leeching also results in decrease in thermal energy storage performance over time due to loss of PCM. As such, there is significant motivation to develop new strategies which would better overcome the challenge of PCM leeching and leakages. In recent years, there has been increased research attention on exploiting an alternate phase transition, namely the solid-solid transition.^24,25 Although these materials have been explored even in the 1960s and 1970s, there have been relatively little efforts toward their practical usage as thermal energy storage material. This can be attributed to their relatively lower latent heat compared to s–l PCMs. However, s–s PCMs have some significant advantage over s–l PCMs. These materials exhibit inherent form stability, completely skirting the issue of leakages and leaching which plague s–l PCMs. Other issues such as potential phase separation from matrix material and volume changes are also likewise avoided.

In general, to prepare s–s PCMs with appreciable latent heat capacity, the material requires a combination of “hard” segments with “soft” segments. During phase transition, the “soft” segment undergoes change in crystallinity, while the “hard” segments help to ensure form stability. Some of the most common examples include PEG (acting as the “soft” segment) being grafted or chemically attached to other support materials (acting as the “hard” segment), such as polyurethane,²⁶ polystyrene,^27–29 cellulose,³⁰ poly(methacrylate),³¹ melamine-formaldehyde networks,³² and even silica.³³ However, of these, with the exception of cellulose, little effort has been made to make use of sustainable materials, with the vast majority of support materials being prepared from petroleum feedstock. To that end, we envisioned that polylactic acid (PLA), a class of biomass-derived and biodegradable polymers,^34–36 can also be used as a sustainable support polymer for PEG-based s–s PCMs. PEG and PLA have been shown to have excellent compatibility,^37,38 with various blends or copolymers being reported for use in drug delivery systems.^39–41 Even though copolymers have been synthezised for various applications,^42–46 PLA–PEG block co-polymers have never been examined for use as s–s PCMs. Furthermore, not every type of PLA–PEG co-polymers would be suitable for thermal energy storage use. For instance, the PLA–PEG block co-polymers reported by Thanomsilp only showed broad DSC peaks.⁴⁷ This is due to the requirement of balancing between the PLA segment and PEG segment so as to ensure form stability while maintaining appreciable specific latent heat.

In this study, we report the preparation PEG methyl ether of 5 kDa (mPEG 5 K) co-polymerized with PLA of different molecular weights to examine the ratio of “hard” to “soft” portion required for form stability. The polymers were characterized by FTIR, NMR, and GPC analysis to confirm the chemical bonding of PLA with PEG. mPEG 5 K physically blended with PLA was also prepared to serve as a control. The form stability and thermal storage capacity of each of these composites/block co-polymers were then tested and evaluated.

EXPERIMENTAL SECTION Materials and instrumentation

Purasorb L (L-lactide) was obtained from Corbion. Diethyl ether, HPLC grade tetrahydrofuran (THF), ε-caprolactone, tin octanoate, and mPEG 5 K were purchased from Sigma Aldrich. Toluene and chloroform were purchased from Fluka. PLA 3051D was purchased from NatureWorks LLC, and deuterated chloroform was purchased from Cambridge Isotope Laboratories. All chemicals were used as received without further purification.

The synthesized mPEG/PLA copolymers were studied by ¹H Nuclear Magnetic Resonance (NMR) (JEOL 500 MHz NMR spectrometer). Approximately 10–20 mg of dried mPEG/PLA copolymer was weighed and dissolved in deuterated chloroform (CDCl₃). The dissolved copolymer was then transferred to an NMR tube for characterization. Gel permeation chromatography (GPC) was analyzed on an Agilent 1260 Infinity II GPC/SEC system equipped with a refractive index detector. THF was used as eluent with a flow rate of 1.0 mL/min at 0°C. Monodisperse poly(styrene) standards were used to obtain the calibration curve. Fourier-transform infrared (FTIR) spectrometry (Vertex 80 v, Bruker) coupled with a diamond attenuated total reflectance (ATR) accessory was performed in the 4000–400 cm⁻¹ wavenumber region, at a spectral resolution of 4 cm⁻¹, and with 64 scans each. Thermogravimetric analysis (TGA) (Q500, TA Instruments) was carried out under a rate of 60 mL/min nitrogen gas flow, with samples of weight range between 5 and 44 mg, and at a heating rate of 20 °C/min from room temperature to 600–900 °C in an alumina crucible. Differential scanning calorimetry (DSC) (Q100, TA Instruments) was performed under a rate of 50 mL/min nitrogen gas flow, with samples of weight range between 2 and 6 mg sealed in aluminum hermetic pans, and at a heating rate of 10 °C/min between a temperature range of 0 °C to 190 °C. Samples were sputtered with a layer of gold before analysis by scanning electron miscopy (SEM) (JSM-6700F, JEOL) using a field-emission scanning electron microscope at a voltage of 5 kV and current at 10 mA.

Synthesis of mPEG/PLA samples via ring opening polymerization (ROP)

The synthesis of the mPEG/PLA polymers were adopted from Jing et al.⁴⁸ with minor modification. The envisaged diblock copolymers mPEG/PLA samples are named mPEG/PLA (x/y) where x and y represent the weight average molecular weight of mPEG and PLA block in kDa, respectively. For instance, to synthesize mPEG/PLA (5 K/2.5 K), 5 g of mPEG (5 K) and 2.5 g of L-lactide were dissolved in 60 mL of toluene in a three-neck flask. Next, 30 mg of tin 2-ethylhexanoate (tin octanoate) was added, and the mixture was refluxed for 2 h at 130 °C under a nitrogen atmosphere. After 2 h, the resultant product was dissolved in sufficient chloroform, stirred vigorously, and precipitated at 0 °C in diethyl ether. The precipitate polymer was filtered and then dried first at room temperature, then under vacuum oven at 60 °C to give a white/grey solid copolymer powder.

Preparation of physically mixed PEG:PLA polymer composite

Commercially available mPEG 5 K and PLA 3051D were physically mixed in mass ratios of 1:2, 1:1 and 2:1 at 190 °C. The melted mixture was then stirred using a spatula until homogeneous and left to cool to room temperature for subsequent characterization. The samples are named PEG:PLA (x:y), where x:y represent their respective ratios.

Form stability and leakage test

A 0.5 g of each respective sample was gently heated to a minimum temperature at which samples were malleable and shaped into a flatten disc like a pellet. The weight of each pellet (W_i) was then recorded and subsequently placed into a petri dish. Leakage test was then conducted, where the samples were placed in a 100 °C oven for 24 h, during which they were taken out at regular timed intervals for form-stability assessment and image capturing, and immediately placed back into the oven. At the end for the leakage test, samples were allowed to cool to room temperature, and the final weight of each pellet was recorded (W_f). The percentage weight change (W_%) of each sample pellet was calculated by W_% = (W_f – W_i)/W_i, and the percentage weight change with respect to PEG (W_%,PEG) in each sample was calculated by W_%,PEG = (W_f – W_i)/W_PEG,Ti, where W_PEG,T is the theoretical initial weight of PEG in each respective composite. The weight of PEG remaining in the sample (W_PEG,r) can be calculated by W_PEG,r = W_PEG,Ti – (W_i – W_f).

RESULTS AND DISCUSSION Structural analysis

¹H NMR spectroscopy of various mPEG/PLA copolymers was shown in Figure 1. The signature peaks in the spectra indicate the successful chemical linking of mPEG copolymer onto the PLA main chain. The peak at around 3.6–3.7 ppm is assigned to the methoxy proton (peak a) and the methylene (–CH₂) (peak b) protons of the mPEG segment. On the other hand, the singlet peak at around 1.5–1.6 ppm is assigned to the methyl protons (–CH₃) present at the PLA (peak d), while the peak between 5.2 and 5.3 ppm is assigned to the methyne proton (–CH) adjacent to the ester functional group of the PLA chains (peak c).^49–52

Figure 1. 1H NMR spectra of (a) mPEG/PLA (5 K/2.5 K), (b) mPEG/PLA (5 K/5 K), and (c) mPEG/PLA (5 K/10 K) copolymer. 1H HMR, nuclear magnetic resonance; mPEG, methoxy-polyethylene glycol; PLA, polylactic acid.

FTIR spectra of the respective polymer composites and starting materials were illustrated in Figure 2. For PLA, the main peaks at 1747 cm⁻¹ were attributed to the stretching vibration of the C = O bond, 1086 cm⁻¹ to C–O stretching, 1452 cm⁻¹ to C–H stretching, and 2997 cm⁻¹ to C–H bending respectively of the PLA segment. Main peaks for PEG include 1095 cm⁻¹ from C–O stretching, 1466 cm⁻¹ from –CH₂ bending, and 2881 cm⁻¹ from C–H stretching respectively.⁵³ FTIR spectra of all PEG:PLA physically mixed composites can be viewed as a simple mathematical overlap of individual PEG and PLA spectra without the emerging of new peaks, except for slight shifts in characteristic peaks. Hence no chemical reaction took place during the physical mixing of the two components. The shift in C = O peak from 1747 cm⁻¹ to 1757 cm⁻¹, shift in C–O–C peak from 1182 cm⁻¹ to 1184–1188 cm⁻¹, and shift in C–H stretching from 1452 cm⁻¹ to 1454 cm⁻¹ could be due to change in chemical environment such as the formation of hydrogen bonding between two components. As for the mPEG/PLA copolymers, when compared with its starting materials, they can be seen that PEG and PLA peaks are present in the polymer's FTIR spectrum and L-lactide peaks have either disappeared or shifted. Since many characteristic peaks of L-lactide are overlapping with the PEG peaks, the main indication of successful polymerization of the PLA segment is that the peak at 933 cm⁻¹ belonging to the –COO lactone ring breathing mode has disappeared for the mPEG/PLA series. Also, L-lactide peaks at 1753 cm⁻¹ belonging to the stretching vibration of the C = O bond has shifted to 1757 cm⁻¹. Peak at 1456 cm⁻¹ due to asymmetric bending in –CH₃ has shifted to 1454 cm⁻¹; and peak at 1265 cm⁻¹ due to asymmetric vibration of the C–O–C bond of the lactone ring has shifted to 1184–1188 cm⁻¹ due to change in chemical environment.⁵⁴ Though a new C–O bond was formed for the mPEG/PLA copolymers, however, due to similar chemical bonds throughout the PLA structure, no distinct new peak was observed.

Figure 2. FTIR analysis of L-lactide, PLA 3051D, mPEG 5 K, and the respective mPEG/PLA diblock polymer and physically mixed composites. FTIR, Fourier-transform infrared; mPEG, methoxy-polyethylene glycol; PLA, polylactic acid.

GPC was further conducted to differentiate between the mPEG/PLA copolymers from the physically mixed PEG:PLA composites (Figure 3). A single peak was observed for the mPEG/PLA copolymers (Figure 3A), suggesting the presence of a single polymer component. This indicates the successful coupling between mPEG and PLA to form the copolymer. The M_n values obtained from GPC for mPEG/PLA(5 K/2.5 K), mPEG/PLA(5 K/5 K), and mPEG/PLA(5 K/10 K) copolymer are 7566, 9170, and 10741 g/mol respectively, which is in good agreement with the theoretical M_n, as shown in Table 1. On the contrary, the physically mixed PLA and PEG showed two distinct peaks (Figure 3B), which suggests PLA and mPEG are not chemical linked and exist as individual components. The M_n value of mPEG is around 7000 g/mol, while the M_n value of PLA is around 65000 g/mol.

Figure 3. GPC curve of (A) copolymer of mPEG/PLA, (B) PEG:PLA physically mixed composites. GPC, gel permeation chromatography; mPEG, methoxy-polyethylene glycol; PLA, polylactic acid.

Table 1 Molecular characterization of mPEG/PLA copolymer and PEG:PLA physically mixed composites.

Samples	Mtheoretical* [g/mol]	Mn [g/mol]	Mw [g/mol]	PDI
mPEG/PLA (5 K/2.5 K)	7500	7566	9401	1.24
mPEG/PLA (5 K/5 K)	10,000	9170	12,663	1.38
mPEG/PLA (5 K/10 K)	15,000	10,741	17,470	1.63
PLA 3051D	-	87,147	150,165	1.72
PEG:PLA (2:1)	-	65,940	95,831	1.45
		7263	8224	1.13
PEG:PLA (1:1)	-	58,146	88,694	1.53
		7114	8078	1.14
PEG:PLA (1:2)	-	67,658	94,391	1.40
		7100	8242	1.16

Abbreviations: mPEG, methoxy-polyethylene glycol; PLA, polylactic acid. *

M_theoretical is the theoretical total molecular weight of mPEG/PLA copolymer by summing each segment.

Form-stability

To test the form stability of the polymer composites, leakage test was performed at 100 °C, and the leakage condition of each sample with respect to time were described in Figure S1. mPEG/PLA (5 K/2.5 K), and PEG:PLA (2:1) and (1:1) showed signs of early form-instability just 1 h into the test, which persisted and left behind an even larger wet patch after 24 h (Figure 4). Though there was no observable wetted area for PEG:PLA (1:2); however, when the pellet was removed from the graph paper at the end of the leakage test, a wetted patch was observed (Figure 4B). Comparing the PEG:PLA physically mixed composites, the wetted area increased with PEG loading, which could be the effect of a lesser amount of PLA matrix to adsorb the PEG to maintain form stability. Leakage occurred for all physically mixed composites, which makes them form-unstable.

Figure 4. Condition of the respective composites (A) at the end of the 24 h leakage test conducted at 100 °C, where red arrows point to the wetted patch. And (B) traced wetted area on graph paper due to leakage. mPEG, methoxy-polyethylene glycol; PLA, polylactic acid.

The mPEG/PLA (5 K/2.5 K) copolymer produced a concentric wetted area, which grew steadily and uniformly throughout the test, in contrast with the physically mixed composites that started with leakage on one side of the sample pellet. This difference in wetting pattern suggests that chemically synthesized PEG/PLA copolymers are more homogeneous than PEG:PLA physically mixed composites which is important for heat absorption or release during the phase transition. In contrast, the other two lower PEG loading mPEG/PLA copolymers (5 K/10 K) and (5 K/5 K) with higher molecular weight PLA components did not show signs of leakage as the PLA molecular weight was sufficiently large to resist dissolvation by PEG. Form stability of a block co-polymer depends on the ratio of “hard” segments (contributed by PLA) to “soft” segment (contributed by PEG). This result suggests that a hard:soft ratio of 1:2 is insufficient to achieve form stability, while ratios of 1:1 or more are sufficient.

Percentage weight change (W_%) of the composites after the leakage test was tabulated (Table S1), and results are consistent with the wetting area where PEG:PLA (2:1) gave the highest weight loss amongst the physically mixed composites, and the other two have almost similar weight change. mPEG/PLA (5 K/2.5 K) showed the highest percentage weight loss among all samples, which was about 1.8 times higher than PEG:PLA (2:1). However, their wetting area appears to be similar in size. This suggests that the material leaked from mPEG/PLA (5 K/2.5 K) might be more viscous due to form-instability instead of pure PEG. mPEG/PLA (5 K/10 K) and (5 K/5 K) on the other hand recorded a considerable percentage mass loss despite no wetting of the graph paper observed. This mass loss could be due to moisture loss from the sample. As such, mPEG/PLA (5 K/10 K) and (5 K/5 K) can be deemed as form-stable.

Phase change properties

The phase change properties of the composites were analyzed by DSC (Figure 5 and Tables S2 and S3. Pure PEG yielded a melting and crystallization enthalpy of 175.20 J/g and 175.50 J/g at 60.03 °C and 40.98 °C respectively. While pure PLA produced an almost negligible melting enthalpy of 1.27 J/g at 155.97 °C with no crystallization peak but a glass transition (T_g) at 60 °C. However, when PLA was chemically bonded to or physically mixed with PEG, a more significant melting and crystallization peak was observed due to the plasticizing effect of PEG.⁵⁵ This plasticization also reduced PLA's crystallinity for both polymer series, and T_m,PLA (melting point of PLA) of the composites were seen to be lower than that of pure PLA.⁵⁶ The decrease in T_m,PLA was especially significant for mPEG/PLA (5 K/2.5 K), which might be due to the fact that it contains the lowest PLA content and low molecular weight. The chemical linkage between the PEG and PLA segment serves as points of defects, and thus mPEG/PLA (5 K/2.5 K) has the largest defect density which severely hinders PLA chain movement and crystallization process. Physically mixed PEG:PLA (2:1) is therefore less affected since they have no chemical linkage to further lock the chains into positions, and have a T_m,PLA similar to the other composites in the series. The PLA enthalpy on the other hand varied proportionally with its weight content for both polymer series.

Figure 5. DSC curves for PLA 3051D, mPEG 5 K and the respective polymer composites. DSC, differential scanning calorimetry; mPEG, methoxy-polyethylene glycol; PLA, polylactic acid.

The latent heat of both series of polymer composites increases linearly with PEG loading. For the physically mixed PEG/PLA composites, as there was no chemical linkage to the matrix to hinder the crystallization of PEG, the latent heat increased proportionately with the increase in the phase change component (PEG) and finally was close to the theoretical values.^57,58 This also resulted in the T_m close to that of mPEG 5 K. However, for the mPEG/PLA copolymers, though the trend was similar, the latent heat was lower than theoretical values by 10%–60%, showing that the lower PEG loading polymer (mPEG/PLA (5 K/10 K)) had a greater deviation. This could be due to two factors: the first is similar to that of the physically mixed composites where the latent heat is proportionate to the overall amount of PEG present in the composites. The second factor is due to PEG being bonded to the PLA matrix, which hinders its full crystallization. Since PLA remains in its solid phase during the PEG's phase transition, the movement of PEG chains will be restricted, leading to the formation of fewer crystalline regions and smaller crystallites, and hence causing the decrease in the enthalpy value and peak melting temperature.^59,60 Therefore as PEG loading decreases, T_m deviates greater from the melting point of pure mPEG 5 K given the larger amount of PLA matrix present that impedes PEG crystallization.

When comparing the phase change properties of the chemically synthesized and physically mixed polymers, the initial latent heat strongly favored the latter, which was found to have minor leakage. Hence, DSC was also reexamined for each composite sample after the leakage test (Tables S2 and S3), and the change in melting enthalpy of PEG component in the mPEG/PLA copolymers after the leakage test (ΔH_%Δ,m,PEG) was found to be at least 5 times smaller than that of the PEG:PLA physically mixed composites (Table S1). mPEG/PLA (5 K/5 K) polymer experienced a 5% drop in ΔH_m,PEG, maintaining a ΔH_{m, PEG} of 52.79 J/g. In comparison, ΔH_m,PEG for mPEG/PLA (5 K/10 K) remained almost unchanged. The trend in the melting enthalpy before and after leakage samples was consistent with the respective amount of PEG remaining within the sample (W_PEG,r). Hence, there is greater phase change reliability and thermal stability for chemically bonded mPEG/PLA copolymers compared with its PEG:PLA physically mixed counterparts.

Thermal cycling of 100 cycles was carried out for the better performing form-stable mPEG/PLA (5 K/5 K) (Figure 6), and the latent heat and peak temperatures were found to be thermally reliable, deviating by about 6% for both ΔH_m,PEG and ΔH_c,PEG, and by 3 °C and 3.7 °C for T_m,PEG and T_c,PEG, respectively. There was no obvious difference in the shape of the DSC curves; hence the stable latent heat and peak temperatures confirmed the thermal reliability of the mPEG/PLA co-polymers for energy storage applications.

Figure 6. Thermal cycling DSC curves mPEG/PLA(5 K/5 K). DSC, differential scanning calorimetry; mPEG, methoxy-polyethylene glycol; PLA, polylactic acid.

As chemically synthesized co-polymers can come in different forms, such as block-co-polymers, alternating co-polymers, graft-co-polymers, and so forth, it was observed from literature that the polymer form does not severely affect the phase change trends and properties. A list of PEG-based block co-polymers and graft co-polymers are summarized in Table 2. Comparing the same block co-polymer structure, PEG_6K-block-PET synthesized by Hu et al.⁶¹ achieved only 39% of its theoretical ΔH_m with a 31% PEG loading, similar to our mPEG/PLA (5 K/10 K) (43% ΔH_m efficiency). While for Li et al.³⁰ who grafted PEG onto cellulose (Cellulose-graft-PEG_2K), when PEG was 50 wt%, the ΔH_m was 67% of the theoretical value, similar to our mPEG/PLA (5 K/5 K) (64% ΔH_m efficiency). When Li et al. increased PEG to 85 wt%, the ΔH_m efficiency also increased to 88%. This rise was similar to the alternating co-block polymer PEG_10K-block-PU(MDI/butanediol fabricated by Su et al.²⁶ with 82% ΔH_m efficiency at 88.7% PEG wt%. Therefore, our block-co-polymer design of mPEG/PLA is not shortchanged due to its polymer form. Nevertheless Sari et al.²⁷ demonstrated that polystyrene-graft-PEG_6K graft-co-polymer achieved a ΔH_m, that is, 157% of its theoretical value. Though there was no discuss on this result, a possible reason could be the phenyl rings along the polystyrene backbone undergoes π-π interactions that help to align the grafted PEG chains to give an overall higher crystallinity. Hence with appropriate design in the polymer structure, they polymer form may enhance the enthalpy.

Table 2 Summary of phase change properties of different reported PEG-based polymers.

[Table omitted. See PDF]

Abbreviation: PEG, polyethylene glycol. *

Phase change properties for PEG_6K was not reported by Hu et al.,⁶¹ hence the value was calculated based on ΔH_m of PEG_6K reported by Sari et al.²⁷

Thermal stability

TGA was performed to study the degradation behavior of PLA 3051D, mPEG 5K and the respective composites (Figure 7). Both PLA 3051D and mPEG 5K decomposed via a one-step degradation, with T_{d, max} at 382 °C and 413 °C, respectively, indicating higher thermal stability for the PEG segment. All other composites decomposed via a two-step degradation. Looking at the percentage weight change of each decomposition step for the respective compounds and comparing it with the theoretical weight percent of each component (Table S4), the first degradation step was assigned to the PLA segment, which involved unzipping depolymerization at the chain ends. The second step was due to the thermal scission of the PEG's main chain.⁶² The mPEG/PLA copolymers exhibited a more distinct two-step degradation, with the PLA segment decomposing at much lower temperatures than pure PLA 3051D (M_n: 87 K, Table 1). This is due to smaller molecular weight of the PLA chain bonded in the copolymers (M_{theoretical,PLA}: 2.5K–10K), which makes it less thermally stable.⁶³ Furthermore, it is likely that trace residual tin catalyst from the synthesis of the polymer also catalyzed the decomposition of PLA resulting in the decomposition temperature being lowered by about 100 °C.⁶⁴ The PEG segment in the copolymer is synthesized using mPEG 5 K, and hence T_d,max for the PEG decomposition falls within a close range. Similarly, the PEG:PLA physically mixed composites used both pure PLA 3051D and mPEG 5 K as starting materials in their mixture, and hence T_d,max for both decomposition step coincides with the respective pure components. The T_98wt% is well above the working temperature range of the PCM composites, thus the form-stable polymers are thermally stable for their intended application.

Figure 7. TGA curves of PLA 3051D, mPEG 5 K and the respective composites. mPEG, methoxy-polyethylene glycol; PLA, polylactic acid; TGA, thermogravimetric analysis.

Surface morphology

SEM was conducted to study the surface morphology of the compounds (Figure 8). PLA 3051D was observed to have a rougher surface than mPEG 5 K. Before the leakage test, for both mPEG/pLA copolymer and PEG:PLA physically mixed composite series, their surface morphology became progressively smoother and formed a continuous layer as PEG loading increased. Surface appeared uniform, indicating good mixing and homogeneity possibly due to the plasticization effect of PEG.

Figure 8. SEM imaging of PLA 3051D, mPEG 5 K and the respective composited before and after leakage test. Scale bar for image (A), (D), (E), (H) is 1 μm, the rest of the image is 10 μm. mPEG, methoxy-polyethylene glycol; PLA, polylactic acid; SEM, scanning electron microscope.

After the leakage test, all composites’ surface morphology increased in roughness and porosity. This could be due to the rearrangement of melted PEG chains around the solid PLA scaffold when heating at 100 °C during the leakage test, or decrease in PEG content due to the leakage, thus exhibiting contoured patterns of the PLA matrix when it cooled. Similarly, samples with a higher PEG loading had a smoother and less porous surface compared with its counterparts after the leakage test. Pores for mPEG/PLA copolymers were observed to be larger than these of the physically mixed composites, possibly due to the fact that PEG bonded to the PLA matrix has limited mobility and thus can only take the shape of the PLA matrix. PEG in the physically mixed composites, however, is unbonded in the melted state and filled the pores, therefore collecting at the bottom. Thus, for the lower PEG loading composites (Figure 8M,N), distinct linear domains were seen protruding on the surface, which could be assigned as PLA domains due to insufficient PEG to cover the PLA surface.

CONCLUSION

Two PCM-matrix systems in the form of chemically bonded mPEG/PLA copolymers and physically mixed PEG:PLA composites were studied. The physically mixed PEG:PLA composites provided a simple fabrication procedure with a higher latent heat recorded. However, they did not attain good form-stability, and their leakage was detected in all three samples. Furthermore, the latent heat for the PEG:PLA composites decreased significantly by 27%–77%. In contrast, the chemical bonded mPEG/PLA copolymers with two lowest PEG loadings exhibited form-stable with the highest latent heat of 55.8 J/g, with very limited drop in comparison with the theoretical value. The chemically bonded mPEG/PLA was more thermally stable and more homogenous as evidenced by TGA analysis and SEM. However, their latent heat could be further improved in our future work, focusing on 1) incorporating different type of PCMs with a higher latent heat such as PEG with high molecular weights or long alkyl chains, and 2) achieving higher crosslinking in the support matrix to retain good form-stability at higher PCM loadings. Nonetheless, this work demonstrated the effectiveness of chemically bonding PCMs to a support matrix to efficiently eliminate leakage problems, and consequently enhanced compatibility between PEG and PLA, providing a useful example for suppressing leakage suffered in PCM applications. On top of that, PLA is bio-sourced and bio-degradable, and PEG is environmentally benign, making our material environmentally sustainable. We are hopeful that with further research, higher latent heat capacities would be achievable for s–s PCMs.

ACKNOWLEDGMENTS

The authors acknowledge the financial support from the Agency for Science, Technology, and Research (A*STAR), Science and Engineering Research Council, and A*ccelerate Technologies for this work (Grant No.: GAP/2019/00314).

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT

Data is available in article's supplementary material. The data that support the findings of this study are available in the supplementary material of this article.